Abstract: Global-scale high temporal and spatial resolution sea surface wind field detection is currently one of the hotspots in global meteorological research and forecasting. In traditional sea surface wind field detection technology, the measurement area is limited, and it is severely restricted by the weather condition. Based on the principle of wind speed retrieval from global navigation satellite system-reflectometry(GNSS-R), the wind speed retrieval model was constructed using the geophysical model function(GMF) method with the level 1 data of BF-1 as input and the wind speed data reanalyzed by the European center for mediumrange weather forecasts (ECMWF) as the reference wind speed. This paper analyzed the influence of different parameters of the level 1 product, such as satellite, observation antenna, specular incident angle, signal-to-noise ratio, etc., on the observation characteristics and wind speed, determined the GMF empirical model parameters and completed the model establishment. The generated retrieved wind speed data and their statistical characteristics were compared and analyzed with the results of cyclone global navigation satellite system(CYGNSS). The retrieved wind speed and the distribution characteristics of the retrieval deviation with respect to the incident angle and the reference wind speed are all consistent with the trend of CYGNSS results. This work preliminarily demonstrates the wind speed detection capability of the BF-1 satellite, which can provide a reference for subsequent detection performance improvement and constellation development.

Key words: global navigation satellite system-reflection, delay-Doppler mapping, geophysical model function, wind speed retrieval, normalized bistatic radar cross-section